Stanford University researchers have recently published a study comparing the performance of ChatGPT-4 with its predecessor, GPT-3.5. The study found that ChatGPT-4, which powers the widely used AI tool ChatGPT, is experiencing a decline in its capabilities. The accuracy for certain tasks has dropped significantly, while GPT-3.5, used in the free version of ChatGPT, has shown improvement. OpenAI’s VP of Product, Peter Welinder, denies the claims, stating that each new version of GPT is smarter than the previous one. The researchers speculate that changes in training data, fine-tuning processes, or usage patterns may be contributing to the decline in performance.